(3) Quantitative Research Flashcards
Questionnaire Design - Contents
- Brand Awareness
- Brand Satisfaction
- Brand Preference
- Purchase Intention
- Personal Info (experience, demographic)
Questionnaire Design - Scales
- MCQ
- Ratings
- Likert scale (strongly agree - strongly disagree)
- Semantic differential scale (poor - good quality)
- Open-ended (not too many!)
Questionnaire Design - Wording
- Be specific (avoid “normally”, “regularly,”…)
- Don’t ask leading questions
- Avoid double-barreled questions (how would you judge X based on Y and Z?)
- Be sensitive to issues of “framing” (alternate frames: should forbid, then, should allow)
Questionnaire Design - Q Order
- Start with non-threatening questions
- Questions should flow logically
- Ask the most important topic first
- Go from broad to specific
- Be aware of order bias and randomize orders
Survey - Administration (3 ways)
Personal
- Instant clarification, high response rate
- Costly, interviewer bias
Telephone
- less costly
- Negative attitude to telemarketing
Mail/Fax/Magazine insert
- Best for personal or embarrassing topics
- Low response rate
Sampling (((())))
Population > All students on campus
Sampling frame > currently enrolled
Sample > 100 students
Element - One student
Sampling Techniques
Random
- every member of the population has aa known probability of being included
Stratified
- Random after grouping the population into strata (per criteria)
Cluster
- Random sampling of groups
Convenience
- easiest population members for the researchers
Random Sampling
- Population
- Assign a number between 1 and N to each element
- Generate n different random numbers between 1 and N
- Form sample
Stratified Random Sampling
- Population
- Grouped into mutually exclusive stata or segments
- Chosen sample forced to contain units from each stratum
Cluster Sampling
- Population
- Divide the population into subgroups
- Select random sample of clusters
- Members of cluster interviewed
Limitations of surveys
- Low response bias
- Causal Relationship
Experiment - Causal Relationship
Condition to establish causality ( X -> Y)
1) Concomitant variation (correlation) : X up/down, Y u/d
2) Time order of occurrence: X before Y
3) Control of alternative factors: X with A -> Y
Experiment - Cause-and-effect
- Measuring cause-and-effect relationships by manipulating independent variables to determine the effect of changes on dependent variables
- Cause-and-Effect Relationship
- IV precedes DV
- Controlling other factors than IV
Experiment - Multiple Factor Design
- More than one Independent factor
- Useful when you believe the effect of one factor depends on the other factor
ex: does the effect of price cuts on sales depend on the level of ad spending ?
Factor 1: Price cut (0% vs. 10%)
Factor 2: Increase in ad spending (0% vs. 50%)